Our initial evaluation of the CBT4CBT program as an adjunct to outpatient treatment indicated statistically significant effects on percent of urine toxicology screens that were positive for illicit drugs as well as significant durability of effects over 6 months. However, despite the enormous promise of computer- based treatments as well as very encouraging initial data from our work and others, well-controlled randomized clinical trials of computer-assisted therapies are still rare. The proposed project will be the first evaluation of CBT4CBT delivered with minimal clinician support, rather than as an adjunct to treatment, and will allow preliminary evaluation of CBT4CBT as a potential 'stand-alone' intervention. It will also be the first trial of CBT4CBT to include a clinician-delivered CBT condition, and thus will permit some exploration of the types of individuals for whom this approach might be most appropriate versus those who respond to clinician-delivered CBT. Specific aims are: To conduct a 12 week randomized trial with 180 treatment-seeking individuals meeting current criteria for illicit drug abuse or dependence. Participants will be randomized to one of three conditions: (1) standard outpatient counseling at a community treatment program, (2) individual clinician delivered CBT, or (3) web-based CBT4CBT with minimal counseling. To evaluate the long-term durability and/or delayed emergence of treatment effects through a six month follow-up after termination of the study treatments. For these analyses, we hypothesize that either version of CBT (clinician and web-based) will be more effective than standard treatment. The primary outcome measures will be reduction in substance use, operationalized as the frequency of substance use by week confirmed by urine toxicology screens. Measures of treatment utilization and relative costs of the interventions will be used to evaluate cost effectiveness of the protocol interventions. As an exploratory aim, we hypothesize that incremental cost-effectiveness ratios (ICERS) will favor computer- delivered over clinician delivered CBT. Other secondary outcomes will be used to evaluate whether web- based CBT4CBT retains key characteristics of traditional clinician-administered CBT (e.g., acquisition of coping skills, use of change strategies) and to evaluate these as potential mediators of outcome as well as several participant variables that may moderate response to clinician-delivered versus computer-delivered CBT.